Multi-resolution transfer learning for rapid prediction of pore-scale multiphase flow

Key Points

• Rapid prediction of pore-scale multiphase flow enhances model efficiency and accuracy, benefiting various applications.
• The approach utilizes transfer learning techniques to reduce the time needed for predictive modeling.
• Analysis leveraged multiple resolutions to simulate multiphase flow dynamics across different scenarios efficiently.
• Findings imply significant improvements in computational methods for complex fluid dynamics, indicating valuable industry applications.